Parent‐Progeny Regression in Indiangrass: Inflation of Heritability Estimates by Environmental Covariances¹

Abstract

Families of two indiangrass [Sorghastrum nutans (L.) Nash] populations were used in this study with a family plot consisting of a parental clone and four half‐sib progeny. The experimental design was a randomized complete block with two replications. The regression of offspring on the parent in the same plot was used to obtain heritability estimates in which environmental covariances might be expected to inflate the parent‐offspring covariance. The regression of offspring in one replication with its parent in another replication, and vice versa, was used to obtain heritability estimates whereby environmental covariances should be minimized. Heritability estimates for heading date and plant height, which had high heritabilities, were similar for the two estimation techniques. Heritability estimates for forage yield and in vitro dry matter digestibility (IVDMD) appeared to be influenced by environmental covariances. Mean heritability estimates for forage yield and IVDMD for the three sets of data analyzed were 0.64 and 0.94 respectively, when parents and offspring in the same plot were used in the regression analyses and 0.28 and 0.42 respectively, when parents and offspring in different plots were used. Heritability estimates for forage protein content averaged 0.49 and were similar for both methods of estimation.

Heritability Estimates for Forage Yield, In Vitro Dry Matter Digestibility, Crude Protein, and Heading Date in Indiangrass1

“…Parent-progeny regression (Equation 9) also was used to obtain heritability estimates. Because environmental covariances caused regression of offspring on parents in the same replication to yield inflated heritability estimates (Vogel et al, 1980a), regression of offspring means in replication 1 on parents in replication 2, and vice versa, were used to obtain heritability estimates with environmental covariances minimized. The regression coefficient is multiplied by 2 to obtain heritability estimates when half-sib means are regressed on parents (Falconer, 1960).…”

Genetic Variation for Selenium Content in Tall Fescue

“…Furthermore, genotype X environmental interactions also influence heritability estimates. An underlying assumption of PO regression is that all genotype X environment interactions between parents and progenies are zero (Hanson, 1963;Dudley and Moll, 1969;Vogel et al, 1980;Casler, 1982). A way to remove this potential bias is to regress progeny means on parental means evaluated under different environments (Buckner et al, 1981).…”

Parent-offspring regression in meadow bromegrass (Bromus riparius Rehm.): Evaluation of two methodologies on heritability estimates

“…If these covariances are positive, the bias to heritability will result in an overly optimistic expectation of genetic advance. Therefore, parents and offspring should be measured in independent environments (Vogel et al 1980;Casler 1982;Nyquist 1991). If the environments are two different years in the same location, they are not independent if a genotype × location interaction exists.…”

“…The first method is to regress progeny means from one environment on parental values from another environment (Vogel et al 1980;Nyquist 1991). This regression can be done in two different ways.…”

“…This variance-covariance method requires that parents and offspring are evaluated together in replicated trials under the same set of environments. The choice of methods would depend on the breeder's specific objectives, either of which would account for potential genotype × environment interactions (Vogel et al 1980;Casler 1982;Nguyen and Sleper 1983;Nyquist 1991).…”